Re: Mach independence

[Ondrej Hurt]

Ondrej Hurt, <address@hidden> wrote:
>
> What about this scenario:
>
> ....................
>

Damn I roughly repeated what had already been written here.
I should had read the mailing list archive :-/

Farid wrote:
>
> Since there is no way to reliably/synchroneously detect the crash
> of a potential IPC sender
> .................

There are exception handlers in L4, why shouldn't they be able to
make one notification RPC ? Because exception thread can get blocked
? Are there other reasons ?

What if the handler got blocked for veeery long, pr-server polled it
with timeout, thought it is dead and then the supposedly dead task
"ressurects" ? I think it should get deleted by the pr-server (or
something else).

Thomas wrote:
>
> But I fear that this kind of polling is going to be expensive.

I think that this polling could be taking place in long intervals
because it would be useful only in rare cases (exception handler got
stuck ?).



______________________________________________________________________
Vyberte Vas "Nejoblibenejsi automobil roku 2002 v CR"
a hlasujte na adrese http://www.autoroku.cz. Hlavni vyhra 600 000 Kc!!!